Help me find the error of my network configs...

[dgeist]

I'm implementing a HA cluster with 3 nodes, ceph, software fencing (PVE 4) and each node has a single "management" network and a pair of 10G fibers to diverse switches with several already-allocated 802.1q vlans.

Here's the logical setup:

eth6 -|
eth9 -|
     bond0-|
          bond0.1000 ("cluster traffic vlan" on the switches)-|
               bond0.1000.10 (intra-cluster corrosync and such)-|   
                    br10 (192.168.10.0/24)

I'm able to manually build all these relationships with what I'd been using previously on an Ubuntu box. I think I'm missing something, though, since now I have to manually associate the br10 with the underlying bond and then "ifconfig up" on everything to get it to start. I eventually want to build similar "sub-vlans" for ceph traffic, etc. This first one is needed before I can setup the cluster feature. Can someone tell me what I'm missing in my network interface config:

auto lo
iface lo inet loopback

auto vmbr0
iface vmbr0 inet static
        <stuff here>
        bridge_ports eth0
        bridge_stp off
        bridge_fd 0

auto br10
iface br10 inet static
   mtu 1500
   address 192.168.10.10
   netmask 255.255.255.0
   network 192.168.10.0
   bridge_ports bond0.1000.10
   bridge_hello 2
   bridge_maxage 12
   bridge_stp off

auto eth6
iface eth6 inet manual

auto eth9
iface eth9 inet manual

auto bond0
iface bond0 inet manual
        slaves eth6 eth9
        mtu 9000
        bond_miimon 100
        bond_mode active-backup

auto bond0.1000
iface bond0.1000 inet manual
   vlan-raw-device bond0
   up /sbin/vconfig add bond0.1000 10

#Q in Q

auto bond0.1000.10
iface bond0.1000.10 inet manual
   vlan-raw-device bond0.1000
   up /sbin/vconfig add bond0 1000

Thanks.
Dan